Lectures           Click on lecture titles to view slides or the buttons to download them as PDFs.

Short Course

Visualizing Model Inference and Robustness  

This is a 9-hour short course version of the full Data Visualization course; the lectures for the full term course are below. Students taking the short course will also need these additional resources:

• Materials for the R review session: A brief introduction to R for data visualization, R code and data for the GDP example. R code and data from the fertility example. Detailed instructions for downloading, installing, and learning my recommended software for quantitative social science are here.
• Materials for Session 3: R code and data for the voting example, and sample output for expected values, first differences, relative risks, and a combination plot. R code and data for the inequality scatterplot, and sample output.
• Materials for Session 5: R code for the crime example.

Topic 1

Course Introduction  

Topic 2

Principles for the Visual Display of Scientific Information  

Topic 3

Cognitive Issues in Visualization  

Topic 4

Graphical Programming in R  

In the first part of the lecture, we will consider examples from ggplot2 collected in this R script, which relies on this dataset.


In the second half of the lecture, time permitting, we will work through an R script that uses grid graphics to solve a basic task, showing confidence intervals around a regression line using this dataset. Finally, a more advanced grid graphics script replaces ticks with gridlines and packs the grid graphics code inside a more general and usable function, contained in this required helper file. The final graphic can be viewed here.


Interested students can find detailed instructions for downloading, installing, and learning my recommended software for quantitative social science here. Focus on steps 1.1 and 1.3 for now, and then, optionally, step 1.2. (Note: These recommendations may seem dated, as many students prefer to use RStudio as an integrated design environment in combination with RMarkdown. You are free to follow that model, which minimizes start-up costs. I still prefer a combination of Emacs, the plain R console, and Latex/XeLatex for my own productivity, with occasional use of Adobe Illustrator for graphics touch-up.)

Topic 5

Exploratory Data Analysis: Between Data & Model    

Topic 6

Visualizing Inference  

Download instructions for the tile package can be found under the Software tab at left. We will discuss up to four examples in detail:

• Making a scatterplot in tile: R code, data, and sample output.
• Visualizing a logit model of voting with tile’s lineplot: R code, data, and sample output for expected values, first differences, relative risks, and a combination plot.
• Making ropeladder plots to show model robustness using crime data: R code.
• Making ropeladder plots to show in-sample simulation results from an ordered probit model of preferenes over carbon taxes from a survey experiment: R code, helper function for in-sample simulation, data, and sample output.

Topic 7

Interactive Visual Displays with R + Shiny

The Shiny package makes it easy to convert your R code and graphics, including those made with the tile package, into interactive displays for the web. We’ll work through this written Shiny tutorial.


We will discuss several Shiny interactives written by your instructor. Feel free to study the applications and code and come to class with questions:

• An interactive scatterplot combining Shiny and ggplot2. This example builds on the running example from the first half of Topic 4 exploring the relationship between life expectancy and GDP per capita in cross-national data. You can download the interactive R code to run locally.
• A simple text-based Shiny interactive with a html-based user interface. This online example tests whether a user-provided sentence is a pangram (a sentence containing every letter in the alphabet). You can also download the underlying R code.
• A more elaborate interactive using Shiny and tile to show who got the most medals in the Olympics using different medal aggregation formulas. The underlying code for the example is in this zip archive.

Topic 8

Advanced Latex for Scientific Typesetting

Time permitting, we will consider the use of modern Latex typesetting tools, especially Xetex and the fontspec package. I offer three stylesheets for students looking to spruce up their documents. (Students new to Latex should read the Not So Short Introduction to Latex before embarking on any of the advanced stylesheets below.)

• caxetexFreeOL (manual). A powerful XeLaTeX stylesheet using free typefaces and implemented for the popular, easy-to-use Latex platform Overleaf. You can find everything you need to get started with caxetexFreeOL at this Overleaf project. Note in particular the template for research papers.
• caxetexFree (manual). The same powerful XeLaTeX stylesheet using free typefaces, but for use on your local computer's TeX installation. You will need to download the relevant typefaces as instructed in the manual.
• caxetexBook (manual). The main XeLaTeX stylesheet I use in my own publishing. You will need to purchase the commercial typefaces listed in the manual if you wish to use this stylesheet.

Gallery 1

Scales and Storytelling  

Gallery 2

Maps as Visual Displays of Information  

Gallery 3

Time Series as Narrative  

See also this excellent confection from XKCD explaining the scale of global temperature variation over the last 20,000 years.

Gallery 4

Grayscale Images of Continuous Data  

Gallery 5

Turning Tables into Graphs  

Gallery 6

Heatmaps for Visualizing Continuous Dyadic Data    

Gallery 7

Ternary Plots for Compositional Data Analysis  

Student Assignments

Problem Set 1  

Due by 25 January 2023; turn in problem 2 early if possible

You will need these data.

Problem Set 2  

Due 22 February 2023

Problem Set 3  

Due 15 March 2023

Breakout Group

Individual memo due before group meets; Group essay due by 27 February

Students will join a small group to discuss a visual display problem of common interest; creation and organization of these groups to be coordinated through the web. Students will write a 2-5 page memo before the first group meeting, and each group will write a 5-8+ page essay for the class on what they have learned, to be distributed by 27 February. Groups will answer questions from the class during the week of 27 February. See the syllabus for further details.

Final Poster

Presented during the final two classes

On an assigned day during the last week of the course, each poster group will present a poster applying the tools learned in class to their own research. Alternatively, students can take an article published in their field and show how better visuals would either more clearly convey the findings or cast doubt on them. The final presentation may address problems raised in the breakout session or problem sets, but it is usually more fruitful for students to tackle a new problem.

Labs

Lab 1

Intro to labs; intermediate R and prediction  

Supplementary material: We will go through this R Markdown file centered on the O-ring example and different methods to compute and visualize confidence intervals. If you want to follow the section while writing your own code, you can use this clean R Markdown file with no code but just notes. See this PDF report directly "knitted" from the R Markdown file, which is one of the recommended ways to produce and submit your homework.

For a crash course in base R and tidyverse, see supplementary slides (also, the source .Rmd; two datasets for practice and knitting: econ.csv and pop.csv).

Lab 2

R Markdown, LaTeX and Overleaf  

Supplementary material: We will go through this R Markdown Sample file, which cover most of its basic functionalities; and you need this JPG to knit the final PDF output.

We will also go through this HW 1 starting file; you need these two figures (Fig 1 and Fig 2) to knit this PDF output.

We will also learn basic LaTeX with Overleaf. We will go through this template, which requires this figure and this bib file to be compiled.

Lab 3

ggplot2  

Supplementary material: Two datasets you need to reproduce the figures for the electric vehicles in Washington State example: ev_data.csv and county_data.csv, with our customized ggplot2 theme. A supplementary .Rmd file containing the main code chunks for reproducing the graphs is available for reference.

Lab 4

advanced ggplot2  

Supplementary material: Four datasets you need to reproduce the figures for the exercises: Nobel prize winners data, Measles data, 92 Presidential election data and Cy Young award data. A supplementary .Rmd file containing the main code chunks for reproducing the graphs is available for reference.

Lab 5

Intro to tile  

Supplementary material: Inequality example's script and data; voting example 's script and data; crime example's script.

Lab 6

Visualizing spatial data  

Supplementary material: To reproduce the graphs, download the Washington Post dataset, and also the shapefiles for the United States (cb_2018_us_state_20m.zip) and New York City. A supplementary .Rmd file containing the main code chunks for reproducing the graphs is available for reference.

Lab 7

Visualizing relational data  

Supplementary material: To reproduce the graphs, download the Medici data and migration flow data. A supplementary .Rmd file containing the main code chunks for reproducing the graphs is available for reference.

Lab 8

Interactive Visual Display with R + Shiny  

Supplementary material: Simple Shiny app script 1, script 2 and a script to demonstrate the use of action button.

Lab 9

Latest extension packages for visualization  

Supplementary material: The underlying .Rmd written with flipbookr + xaringan packages.